PERFORMING DE-DUPLICATION FOR AT LEAST ONE COMPUTER FILE IN A COMPUTER SYSTEM

Abstract

The present invention provides a method and system of performing de-duplication for at least one computer file in a computer system. In an exemplary embodiment, the method and system include (1) tuning a rolling-hash algorithm for the de-duplication, (2) chunking the data in the file into chunks of data by using the tuned algorithm, (3) producing a content identifier for each of the chunks, and (4) processing the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique. In an exemplary embodiment, the computer system includes a de-duplication-enabled data store. In an exemplary embodiment, the computer system includes (a) a transferor computer system that is configured to transfer the file to a de-duplication-enabled computer system and (b) the de-duplication-enabled computer system.

Description

BACKGROUND OF THE INVENTION

1. Problems with De-Duplicated File-Systems

Computer systems (e.g. server computer systems) need the ability to perform efficient data de-duplication on data. Backup solutions for computer file-system have been including some form of data “de-duplication” or data “redundancy elimination” algorithms. These algorithms can be used at the whole-file or at the sub-file level.

2. Prior Art Systems

Referring to FIG. 1, a common prior art system for performing sub-file de-duplication is to first break data streams (files) into chunks using a rolling-hash algorithm, such as Rabin fingerprinting. Rolling-hash fingerprinting algorithms can be set to produce chunks of an “expected size” based on parameters of the algorithm. Once the files are in chunks, a collision-resistant hashing algorithm is used to uniquely identify the content of each of those chunks by generating unique identifiers.

These unique identifiers are then placed into an index that can be queried. When a chunk is found which already exists in the system (found by querying the index or attempting an insert and getting a collision), that chunk can be replaced by a reference to that chunk, and “de-duplication” occurs. For each file that is chunked, a “blueprint” (references to unique chunks) is produced identifying how to reconstruct the file from its constituent parts or chunks.

The size of the data chunks has a dramatic effect on the de-duplication rates. As the size of the chunks decreases, the de-duplication rate increases, and the size of the unique content identifier index and the size of the “blueprints” used to reconstruct objects from their parts increase, resulting in slower object “reconstruction” from the “blueprints”. As the size of the chunks increases, the size of the unique content identifier index and the size of the “blueprints” decrease while the object “reconstruction” from the “blueprints” becomes faster.

One problem with this prior art de-duplication system relates to certain types of files (e.g., JPEGs, MPEGs, other similarly compressed media files) that are very unlikely to have duplicate sub-file chunks with other files. Specifically, these types of files can use a very large chunk size, whereas other files in the system can use a smaller chunk size. Other factors may also influence whether a certain file type is a candidate for smaller or larger chunking. For example, file types that are typically “consumed” by the user(s) of the system containing these files are generally poor candidates for data de-duplication, and, thus, would benefit from larger block or whole-file de-duplication.

File types that are typically “produced” by the user(s) of the system containing those files would typically be good candidates for data de-duplication. This is because such files will often be found in different stages of preparedness (e.g., drafts, modifications, alterations). For example, movie files (such as those available from YouTube, LLC) often have multiple versions of the same movie and would benefit from smaller-block data de-duplication.

A solution is required which can achieve high speeds without sacrificing de-duplication rates. Therefore, a method and system of performing de-duplication for at least one computer file in a computer system is needed.

SUMMARY OF THE INVENTION

The present invention provides a method and system of performing de-duplication for at least one computer file in a computer system. In an exemplary embodiment, the method and system include (1) tuning a rolling-hash algorithm for the de-duplication, (2) chunking the data in the file into chunks of data by using the tuned algorithm, (3) producing a content identifier for each of the chunks, and (4) processing the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique. In an exemplary embodiment, the computer system includes a de-duplication-enabled data store. In an exemplary embodiment, the computer system includes (a) a transferor computer system that is configured to transfer the file to a de-duplication-enabled computer system and (b) the de-duplication-enabled computer system.

In an exemplary embodiment, the tuning includes tuning the algorithm based on the type of content in the file. In an exemplary embodiment, the tuning includes tuning the algorithm based on the size of the file. In an exemplary embodiment, the tuning includes tuning the algorithm based on the resource constraints of the system. In a specific embodiment, the tuning includes tuning the algorithm based on the remaining space in a data store of the system. In a specific embodiment, the tuning includes tuning the algorithm based on the bandwidth of a transmission medium of the system.

In an exemplary embodiment, the tuning includes tuning the algorithm based on file type usage in the system. In a specific embodiment, the tuning includes, if the file type usage is produced, decreasing a chunk size that the algorithm will produce. In a specific embodiment, the tuning includes, if the file type usage is consumed, increasing a chunk size that the algorithm will produce.

In an exemplary embodiment, the tuning includes adjusting a minimum chunk size that the algorithm will produce. In an exemplary embodiment, the tuning includes adjusting an average expected chunk size that the algorithm will produce. In an exemplary embodiment, the tuning includes adjusting a maximum chunk size that the algorithm will produce. In an exemplary embodiment, the tuning includes adjusting an input window size upon which the algorithm will operate.

In an exemplary embodiment, the producing includes applying a collision-resistant hashing algorithm to each of the chunks. In a specific embodiment, the applying includes applying SHA-1 to each of the chunks.

In an exemplary embodiment, the processing includes, if an index of the computer system does not include the content identifier, (a) identifying the chunk corresponding to the content identifier as being a unique chunk, (b) identifying the content identifier as a unique content identifier, and (c) adding the unique content identifier to the index. In an exemplary embodiment, the processing includes, if an index of the computer system includes the content identifier, (a) identifying the chunk corresponding to the content identifier as being a duplicated chunk, (b) removing the duplicated chunk from the system, and (c) adding a reference to the identified chunk.

In an exemplary embodiment, the processing includes, if an index of the computer system does not include the content identifier, (a) identifying the chunk corresponding to the content identifier as being a unique chunk, (b) receiving the identified chunk, (c) identifying the content identifier as a unique content identifier, and (d) adding the unique content identifier to the index. In an exemplary embodiment, the processing includes, if an index of the computer system includes the content identifier, (a) identifying the chunk corresponding to the content identifier as being a duplicated chunk and (b) adding a reference to the identified chunk.

The present invention also provides a computer program product usable with a programmable computer having readable program code embodied therein of performing de-duplication for at least one computer file in a computer system. In an exemplary embodiment, the computer program product includes (1) computer readable code for tuning a rolling-hash algorithm for the de-duplication, (2) computer readable code for chunking the data in the file into chunks of data by using the tuned algorithm, (3) computer readable code for producing a content identifier for each of the chunks, and (4) computer readable code for processing the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique.

THE FIGURES

FIG. 1 is a flowchart of a prior art technique.

FIG. 2A is a flowchart in accordance with an exemplary embodiment of the present invention.

FIG. 2B is a diagram in accordance with an exemplary embodiment of the present invention.

FIG. 3A is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 3B is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 3C is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 3D is a flowchart of the tuning step in accordance with a specific embodiment of the present invention.

FIG. 3E is a flowchart of the tuning step in accordance with a specific embodiment of the present invention.

FIG. 3F is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 3G is a flowchart of the tuning step in accordance with a specific embodiment of the present invention.

FIG. 3H is a flowchart of the tuning step in accordance with a specific embodiment of the present invention.

FIG. 3I is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 3J is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 3K is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 3L is a flowchart of the tuning step in accordance with an exemplary embodiment of the present invention.

FIG. 4A is a flowchart of the producing step in accordance with an exemplary embodiment of the present invention.

FIG. 4B is a flowchart of the applying step in accordance with a specific embodiment of the present invention.

FIG. 5A is a flowchart of the processing step in accordance with an exemplary embodiment of the present invention.

FIG. 5B is a flowchart of the processing step in accordance with an exemplary embodiment of the present invention.

FIG. 5C is a flowchart of the processing step in accordance with an exemplary embodiment of the present invention.

FIG. 5D is a flowchart of the processing step in accordance with an exemplary embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a method and system of performing de-duplication for at least one computer file in a computer system. In an exemplary embodiment, the method and system include (1) tuning a rolling-hash algorithm for the de-duplication, (2) chunking the data in the file into chunks of data by using the tuned algorithm, (3) producing a content identifier for each of the chunks, and (4) processing the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique. In an exemplary embodiment, the computer system includes a de-duplication-enabled data store. In an exemplary embodiment, the computer system includes (a) a transferor computer system that is configured to transfer the file to a de-duplication-enabled computer system and (b) the de-duplication-enabled computer system.

Referring to FIG. 2A, in an exemplary embodiment, the present invention includes a step 212 of tuning a rolling-hash algorithm for the de-duplication, a step 214 of chunking the data in the file into chunks of data by using the tuned algorithm, a step 216 of producing a content identifier for each of the chunks, and a step 218 of processing the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique. Referring to FIG. 2B, in an exemplary embodiment, the present invention includes a tuner 222, a chunker 224, and a content identifier producer 226.

In an exemplary embodiment, tuner 222 tunes the rolling-hash algorithm 232 based attributes of the computer file 230 and system properties 231 (e.g., resource constraints of the system). In an exemplary embodiment, Chunker 224 then chunks computer file 230 into chunks, such as chunk 234 and chunk 236, via tuned rolling-hash algorithm 232. In an exemplary embodiment, Content Identifier Producer 226 thereafter produces a content identifier for each of the chunks, content identifier 238 for chunk 234 and content identifier 239 for chunk 236. In an exemplary embodiment, the present invention, via step 218, then processes the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique, via a De-duplicator 240, an Index of Unique Chunks 242, a References to Unique Chunks 244 (e.g., a “blueprint”), and a Data Store 246.

Tuning the Algorithm

Referring to FIG. 3A, in an exemplary embodiment, tuning step 212 includes a step 312 of tuning the algorithm based on the type of content in the file. Referring to FIG. 3B, in an exemplary embodiment, tuning step 212 includes a step 314 of tuning the algorithm based on the size of the file. Referring to FIG. 3C, in an exemplary embodiment, tuning step 212 includes a step 316 of tuning the algorithm based the resource constraints of the system. Referring to FIG. 3D, in a specific embodiment, tuning step 316 includes a step 318 of tuning the algorithm based on the remaining space in a data store of the system. Referring to FIG. 3E, in a specific embodiment, tuning step 316 includes a step 319 of tuning the algorithm based on the bandwidth of a transmission medium of the system.

Referring to FIG. 3F, in an exemplary embodiment, tuning step 212 includes a step 322 of tuning the algorithm based on file type usage in the system. Referring to FIG. 3G, in a specific embodiment, tuning step 322 includes a step 324 of, if the file type usage is produced, decreasing a chunk size that the algorithm will produce. Referring to FIG. 3H, in a specific embodiment, tuning step 322 includes a step 326 of, if the file type usage is consumed, increasing a chunk size that the algorithm will produce. Referring to FIG. 3I, in an exemplary embodiment, tuning step 212 includes a step 332 of adjusting a minimum chunk size that the algorithm will produce. The minimum chunk size of the algorithm can prevent the chunks from being too small. Referring to FIG. 3J, in an exemplary embodiment, tuning step 212 includes a step 334 of adjusting an average expected chunk size that the algorithm will produce. The average expected chunk size of the algorithm can be tuned by adjusting the size of the data fingerprint mask of the algorithm. A bigger mask leads to a larger expected chunk size, while a smaller mask leads a smaller expected chunk size. In an exemplary embodiment, the tuning includes running a modulus operation against the rolling hash value for finer-grained control of the average expected chunk size.

Referring to FIG. 3K, in an exemplary embodiment, tuning step 212 includes a step 336 of adjusting a maximum chunk size that the algorithm will produce. The maximum chunk size of the algorithm can prevent chunks from being too large. Referring to FIG. 3L, in an exemplary embodiment, tuning step 212 includes a step 338 of adjusting an input window size upon which the algorithm will operate. The input window size is the number of bytes in the byte-range that the rolling-hash algorithm examines in order to produce a fingerprint for a given offset in the file.

In an exemplary embodiment, the tuning includes setting the minimum chunk size, the average expected chunk size, and the maximum chunk size to the same value. This setting could constitute fixed-block chunking. In an exemplary embodiment, the tuning includes setting the minimum chunk size, the average expected chunk size, the maximum chunk size, and the maximum chunk size to their maximum values. This setting could constitute whole-file de-duplication.

Producing a Content Identifier

Referring to FIG. 4A, in an exemplary embodiment, producing step 216 includes a step 412 of applying a collision-resistant hashing algorithm to each of the chunks. Referring to FIG. 4B, in a specific embodiment, applying step 412 includes a step 414 of applying SHA-1 to each of the chunks.

Processing

Referring to FIG. 5A, in an exemplary embodiment, processing step 218 includes, if an index of the computer system does not include the content identifier, a step 512 of identifying the chunk corresponding to the content identifier as being a unique chunk, a step 514 of identifying the content identifier as a unique content identifier, and a step 516 of adding the unique content identifier to the index. Referring to FIG. 5B, in an exemplary embodiment, processing step 218 includes, if an index of the computer system includes the content identifier, a step 522 of identifying the chunk corresponding to the content identifier as being a duplicated chunk, a step 524 of removing the duplicated chunk from the system, and a step 526 of adding a reference to the identified chunk.

Referring to FIG. 5C, in an exemplary embodiment, processing step 218 includes, if an index of the computer system does not include the content identifier, a step 532 of identifying the chunk corresponding to the content identifier as being a unique chunk, a step 534 of receiving the identified chunk, a step 536 of identifying the content identifier as a unique content identifier, and a step 538 of adding the unique content identifier to the index. Referring to FIG. 5D, in an exemplary embodiment, processing step 218 includes, if an index of the computer system includes the content identifier, a step 542 of identifying the chunk corresponding to the content identifier as being a duplicated chunk and a step 544 of adding a reference to the identified chunk.

General

The present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment containing both hardware and software elements. In an exemplary embodiment, the present invention is implemented in software, which includes but is not limited to firmware, resident software, and microcode.

Furthermore, the present invention can take the form of a computer program product accessible from a computer-usable or computer-readable medium providing program code for use by or in connection with a computer system or any instruction execution system. The computer program product includes the instructions that implement the method of the present invention. A computer-usable or computer readable medium can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The medium can be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device) or a propagation medium. Examples of a computer-readable medium include a semiconductor or solid-state memory, magnetic tape, a removable computer diskette, a random access memory (RAM), a read-only memory (ROM), a rigid magnetic disk, and an optical disk. Current examples of optical disks include compact disk-read only memory (CD-ROM), compact disk-read/write (CD-R/W), and DVD.

A computer system suitable for storing and/or executing program code includes at least one processor coupled directly or indirectly to memory elements through a system bus. The memory elements include local memory employed during actual execution of the program code, bulk storage, and cache memories that provide temporary storage of at least some program code to reduce the number of times code is retrieved from bulk storage during execution. Input/output (I/O) devices (including but not limited to keyboards, displays, pointing devices, etc.) can be coupled to the computer system either directly or through intervening I/O controllers. Network adapters may also be coupled to the computer system in order to enable the computer system to become coupled to other computer systems or remote printers or storage devices through intervening private or public networks. Modems, cable modems, and Ethernet cards are just a few of the currently available types of network adapters. The computer system can also include an operating system and a compute file-system.

CONCLUSION

Having fully described a preferred embodiment of the invention and various alternatives, those skilled in the art will recognize, given the teachings herein, that numerous alternatives and equivalents exist which do not depart from the invention. It is therefore intended that the invention not be limited by the foregoing description, but only by the appended claims.

Claims

1. A method of performing de-duplication for at least one computer file in a computer system, the method comprising:

tuning a rolling-hash algorithm for the de-duplication;

chunking the data in the file into chunks of data by using the tuned algorithm;

producing a content identifier for each of the chunks; and

processing the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique.

2. The method of claim 1 wherein the tuning comprises tuning the algorithm based on the type of content in the file.

3. The method of claim 1 wherein the tuning comprises tuning the algorithm based on the size of the file.

4. The method of claim 1 wherein the tuning comprises tuning the algorithm based on the resource constraints of the system.

5. The method of claim 4 wherein the tuning comprises tuning the algorithm based on the remaining space in a data store of the system.

6. The method of claim 4 wherein the tuning comprises tuning the algorithm based on the bandwidth of a transmission medium of the system.

7. The method of claim 1 wherein the tuning comprises tuning the algorithm based on file type usage in the system.

8. The method of claim 7 wherein the tuning comprises, if the file type usage is produced, decreasing a chunk size that the algorithm will produce.

9. The method of claim 7 wherein the tuning comprises, if the file type usage is consumed, increasing a chunk size that the algorithm will produce.

10. The method of claim 1 wherein the tuning comprises adjusting a minimum chunk size that the algorithm will produce.

11. The method of claim 1 wherein the tuning comprises adjusting an average expected chunk size that the algorithm will produce.

12. The method of claim 1 wherein the tuning comprises adjusting a maximum chunk size that the algorithm will produce.

13. The method of claim 1 wherein the tuning comprises adjusting an input window size upon which the algorithm will operate.

14. The method of claim 1 wherein the producing comprises applying a collision-resistant hashing algorithm to each of the chunks.

15. The method of claim 14 wherein the applying comprises applying SHA-1 to each of the chunks.

16. The method of claim 1 wherein the processing comprises, if an index of the computer system does not include the content identifier:

identifying the chunk corresponding to the content identifier as being a unique chunk;

identifying the content identifier as a unique content identifier; and

adding the unique content identifier to the index.

17. The method of claim 1 wherein the processing comprises, if an index of the computer system includes the content identifier:

identifying the chunk corresponding to the content identifier as being a duplicated chunk;

removing the duplicated chunk from the system; and

adding a reference to the identified chunk.

18. The method of claim 1 wherein the processing comprises, if an index of the computer system does not include the content identifier:

identifying the chunk corresponding to the content identifier as being a unique chunk;

receiving the identified chunk;

identifying the content identifier as a unique content identifier; and

adding the unique content identifier to the index.

19. The method of claim 1 wherein the processing comprises, if an index of the computer system includes the content identifier:

identifying the chunk corresponding to the content identifier as being a duplicated chunk; and

adding a reference to the identified chunk.

20. A system of performing de-duplication for at least one computer file in a computer system, the system comprising:

a tuning module configured to tune a rolling-hash algorithm for the de-duplication;

a chunking module configured to chunk the data in the file into chunks of data by using the tuned algorithm;

a producing module configured to produce a content identifier for each of the chunks; and

a processing module configured to process the chunks that are unique, the content identifier for each of the chunks that are unique, and references to the chunks that are unique.

21. The system of claim 20 wherein the tuning module comprises a tuning module configured to tune the algorithm based on the type of content in the file.

22. The system of claim 20 wherein the tuning module comprises a tuning module configured to tune the algorithm based on the size of the file.

23. The system of claim 20 wherein the tuning module comprises a tuning module configured to tune the algorithm based on the resource constraints of the system.

24. The system of claim 23 wherein the tuning module comprises a tuning module configured to tune the algorithm based on the remaining space in a data store of the system.

25. The system of claim 23 wherein the tuning module comprises a tuning module configured to tune the algorithm based on the bandwidth of a transmission medium of the system.

26. The system of claim 20 wherein the tuning module comprises a tuning module configured to tune the algorithm based on file type usage in the system.

27. The system of claim 26 wherein the tuning module comprises, if the file type usage is produced, a decreasing module configured to decrease a chunk size that the algorithm will produce.

28. The system of claim 26 wherein the tuning module comprises, if the file type usage is consumed, an increasing module configured to increase a chunk size that the algorithm will produce.